From here to there

Connections

first step

Connecting Modules - 1

first step

Connecting Modules - 2

first step

Connecting Modules - 3

  • OSMOSE experience
  • The other OSMOSE experience
  • The optimizer experience
    • Flat file connections

Biology on Poseidon’s end

first step

Biology on Poseidon’s end - 2

first step

Biology on Poseidon’s end - 3

  • Event Based
  • Order is important
  • Reactive

Simple get biomass

first step

OSMOSE get biomass

first step

Object Oriented Fishing

first step

Object Oriented Fishing

  • Where is the fish ?
    • LocalBiology
  • Fish movement/growth?
    • Events modifying LocalBiology
    • Or entirely exogenous
  • What decides how many catches?
    • Gear object

Simple React

first step

OSMOSE React

first step

Slices

  • Build a model not a cake
  • Sources of opacity
    • Opaque code
    • Complexity
  • Slices
    • Simplest geography, biology, fishery
    • Iteratively make complex

Policies

Size premiums

  • Market prices
    • by species
    • by port
    • by time
    • by size
  • Change in hold size information
  • Inference overhaul

MSC prices for size incentives - 2

Delays in licensing

  • Very little new code
  • Results may be obvious

Kitchen-sink - 1

first step

Kitchen-sink - 2

first step

Kitchen-sink - 3

first step

Delays in unloading fish

  • Easy as long as we know its effect on sale prices
  • Results may be obvious

Adaptive Effort Regulations

Adaptive Effort Regulations - 2

  • Straighforward application of Reinforcement Learning
  • Might start initially with some pre-made rules

Partial Spatial Effort Regulation

  • Attach additional spatial cost
  • Might be confusing if effort prices change fast

Partial Spatial Effort Regulation - 2

Opt-in Gear

  • Easy to implement
  • Hard to get transition speed right

Opt-in Gear - 2

Market Concentration

  • Completely separate model ?
  • Can we just grab some IO model?

Cheating/Different regimes

  • All already coded in

Fishers

first step

Change Base

  • Moving as a discrete choice : easy
  • Long-term building of new ports: hard

Translocation

  • ‘Threshold’ translocation is easy to add
  • Might be too “effective” without much gear noise

Discarding

  • Already coded in terms of weight
  • Needs to be made size/age specific

Entry - Exit

  • Easy to implement
  • Hard to get transition speed right
  • Usually very strong effects

Habitat Damage

  • Easy to implement
  • Needs good feedback loop into the biology layer

Biology

Age/Length Structure

  • Already in
  • Missing habitat effects
  • Missing predator/prey effects

Depth Layers

  • Major work
    • Easier on the decision side
    • Hard on visualization
    • Harder on the biology side